Mitotic role for the Cdc28 protein kinase of Saccharomyces cerevisiae

Transcription

1 Proc. Nati. cad. Sci. US Vol. 87, pp , ugust 1990 Genetics Mitotic role for the Cdc28 protein kinase of Saccharomyces cerevisiae STEVEN I. REED ND CURT WITTENBERG Department of Molecular Biology, MB-7, Research Institute of Scripps Clinic, North Torrey Pines Road, La Jolla, C Communicated by Ira Herskowitz, May 21, 1990 (received for review February 21, 1990) BSTRCT The Cdc28 protein kinase functions in the G, to S phase transition of the cell cycle of the budding yeast Saccharomyces cerevisiae. This is in contrast with observations of the homologous protein kinase from a variety of metazoans, where activity and function are associated with the G2 to M phase transition. We present evidence that the Cdc28 protein kinase is also required for mitosis and that this function is executed in the G2 interval of the cell cycle. We show, in addition, that the protein kinase is highly active during this phase of the cell cycle. The dual role ofthe Cdc28 protein kinase in the S. cerevisiae cell cycle thus parallels that demonstrated for the cdc2 protein kinase of the fission yeast Schizosaccharomyces pombe. The characterization of a highly conserved protein kinase in a wide variety of eukaryotic organisms has led to the convergence of a number of active areas of biological investigation. Members of the Cdc28/cdc2 protein kinase family, defined initially by mutational analysis of the cell cycles of budding and fission yeasts, respectively, are associated with cell cycle transitions in mitosis and meiosis (1-4). Parallel to the genetic analyses, most biochemical and physiological work has implicated this kinase in the induction of the mitotic state or in the progression from meiotic prophase to metaphase. In somatic cells and cleavage embryos, kinase activity has been shown to peak at the G2/M phase boundary (5-7). The same active protein kinase has been shown to induce the maturation of oocytes in marine invertebrates and frogs, where it is a component of maturation-promoting factor (8-12). In all of these cases, the active enzyme is a heteromultimeric protein of imprecisely determined stoichiometry composed of a 34-kDa catalytic subunit and 45- to 60-kDa regulatory subunit known as cyclin and, under some circumstances, other polypeptides (6, 7, 13-18). The mitotic and meiotic activities of the kinase, usually assayed by its ability to phosphorylate histone H1 in vitro, correlate with the assembly of these protein complexes as well as with covalent modifications of the catalytic subunit, specifically tyrosine and threonine dephosphorylations (5, 11, 19-22). The activity of the kinase decays rapidly as cells complete mitosis and there is some evidence that this decay is essential to reestablish interphase (5-7, 14, 23). The catastrophic proteolysis of cyclin is thought to cause or facilitate this inactivation of the kinase (6, 24). In contrast to the assignment of a mitotic role through biochemical studies, a role in the G1 to S phase transition has been established through genetic analysis in budding and fission yeasts, Saccharomyces cerevisiae and Schizosaccharomyces pombe, respectively (1-3). In fact, in budding yeast, most analysis has concentrated on the G1 function of the Cdc28 protein kinase since the standard mutant alleles utilized do not confer a significant defect in mitotic functions (1, 2). In budding yeast, an active Cdc28 protein kinase complex has been characterized, the presence ofwhich correlates with The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C solely to indicate this fact a cell's ability to undergo the transition from G1 to S phase (15, 25). External signals that cause cells to arrest in GI phase lead to dissociation of the active complexes (15). lthough the precise composition and stoichiometry of G1 complexes remain to be determined, there is some evidence that cyclinlike proteins perform an analogous function to that of cyclins in the G2 to M phase transition (26-29). The products of the genes CLNI, CLN2, and CLN3 are polypeptides with limited primary structure similarity to mitotic cyclins (26, 27). These have been shown to associate with the CDC28 gene product in G1 (45) and genetic evidence suggests that they are rate-limiting for the GI to S phase transition and required for active GI Cdc28 protein kinase complexes (26-29). Presumably, a similar regulatory system exists in fission yeast where the cdc2 protein kinase is also required for the G1 to S phase transition (3), although no GI activity has been detected (17, 18). It is not known whether the homologous protein kinase from animal cells also has a G1 role. However, Cdc2Hs, the human homolog, can perform the GI function in budding and fission yeast (30, 31). mitotic role for the Cdc28 protein kinase has been reported (32). Piggott et al. (32) described a mutant allele, cdc28-jn, which they claimed arrested with a homogeneous G2 cell division cycle phenotype. They also reported that other recessive cdc28 mutations conferred a G2 phase cell cycle defect that was normally not observed because the execution point for this function was in late G1, not well separated from the earlier execution point associated with G1 arrest (32). Most cells in an asynchronous culture would be expected to exhibit arrest at the first of two closely spaced CDC28-dependent events. Both of the genetic results reported by Piggott et al. (32) have been difficult to reproduce (see Discussion). On the other hand, the Cdc28 protein kinase is active when cells are arrested in G2 or M phase using mutants and inhibitors (15, 25, 33). Furthermore, the Cdc28 protein kinase can perform mitotic functions when expressed in fission yeast in place of the Cdc2 protein kinase (34).These observations and the high degree of conservation of mitotic control through the eukaryotic phylogeny led us to reinvestigate the possibility of a mitotic role for the Cdc28 protein kinase in Sa. cerevisiae. We report here that based on genetic and biochemical criteria, the Cdc28 protein kinase plays a role in mitotic induction analogous to that of the cdc2 protein kinase of fission yeast and of its homologs in animal cells. MTERILS ND METHODS Strains and Growth Media. ll yeast strains were congenic with BF264-1SD, which has been described (35). dditional markers were introduced from plasmids by transformation (36, 37). Standard culture media for Sa. cerevisiae were used in all experiments except as specified (38). Conditional Shift Experiments. Wild-type, cdc284, and cdc28-13 cells were grown in rich medium (YEPD) to a concentration of 2-4 x 101 cells per ml at 23TC at which time hydroxyurea (39) was added to 0.2 M. Incubation at 230C continued for 4 hr and then the culture was shifted to 380C. fter 1 hr at 380C, cells were collected on 2.5-cm membranes (0.45-gm pore size) using a Millipore filtration apparatus and

2 5698 Genetics: Reed and Wittenberg quickly resuspended in fresh medium without hydroxyurea at 380C. Incubation was continued under these conditions for the remainder of the experiment. s specified, samples were removed and fixed for microscopic analysis and for flow cytometric analysis. To arrest CDC28 and cdc28 mutant cells in G2 phase, isogenic derivatives were prepared in which the MIHJ (40) gene was deleted and a copy of the Sc. pombe wee] mitotic inhibitor gene under control of the regulatable Sa. cerevisiae GL) promoter was integrated (40). Cells were grown in rich medium containing the neutral carbon source raffinose (YEPraf) at 230C. Galactose was added to 2% (wt/vol) and cells were grown at 230C for 3.5 hr at which time most cells had a distinctive morphology associated with G2 arrest in the context of an mihl/weel block (40). Cells were shifted to 380C for 0.5 hr and then filtered and resuspended in prewarmed YEPD (repressing) medium. Incubation for the remainder of the experiment was at 380C or at 360C. Microscopy and Flow Cytometry. For determination of cell number and budding index, cells were fixed in 3.7% (vol/vol) formaldehyde. fter fixation overnight, cell aggregates were disrupted with a 20-sec pulse using a Braun sonicator and a needle probe. Leitz SM-Lux phase-contrast microscope fitted with a x40 objective and a hemacytometer were used for counting and determination of budding index (percentage of budded cells). For fluorescence microscopy and flow cytometry, cells were stained with propidium iodide as described (41). Flow cytometric analysis using a FCS IV analyzer (Becton Dickinson) has been described (41). For analyses where cell volume increased dramatically during the course of the experiment after an extended cell cycle arrest, an upward drift of peaks was unavoidable due to increasing cytoplasmic autofluorescence. Fluorescence microscopy of propidium iodide-stained cells was performed using a Zeiss xiophot photomicroscope fitted with a X 100 objective. combination of fluorescence and phase-contrast optics was used to accentuate cellular contours simultaneously with nuclear staining. Tri-X film (Kodak) was used. Determination of Cdc28 Protein Kinase ctivity. Cells were synchronized either by arrest and subsequent release using mating pheromone or by zonal elutriation. Synchronization by the mating pheromone has been described (25). For synchronization by centrifugal elutriation, 1 liter of cells growing in minimal sucrose medium (YNBsuc) at 5 x 107 cells per ml was loaded on a Beckman JE1OX rotor at 4200 rpm and a flow rate of 35 ml/min. Small unbudded presumably newly released daughter cells were obtained by increasing the flow rate to 40 ml/min. mixed population containing larger and budded cells was obtained by stepping the flow rate to 55 ml/min. Cdc28 protein kinase assays were performed essentially as described (15, 25). RESULTS Cdc28 Protein Kinase Is Required fter an S-Phase Block to Induce Mitosis. When temperature-sensitive cdc28 mutants were arrested in S phase using the inhibitor hydroxyurea and then shifted to the restrictive temperature at the time of removal of the inhibitor, most cells progressed from S phase through mitosis and were arrested in G1 (unpublished results). These observations suggest that the Cdc28 protein kinase performs no essential role late in the cell cycle. This interpretation, however, is difficult to reconcile with a large body of experimental observations in other systems. possible explanation for the failure to detect a mitotic role for Cdc28 is that the mitotic demand for kinase activity may be lower than that in G1 coupled with a significant time requirement for decay of mutant Cdc28 protein. lternatively, the G1 forms of the temperature-sensitive mutant kinase may be Proc. Natl. cad. Sci. US 87 (1990) more thermolabile than the mitotic forms. Either of these might cause the G1 function to become limiting first in a population of asynchronous or synchronous mutant cells shifted to the restrictive temperature. We therefore modified the experiment described above by using a more extreme restrictive temperature than is normally employed to facilitate the decay of the thermolabile mutant protein. In addition, the mutant cells were preincubated at the restrictive temperature for 1 hr after synchronization but prior to removal of the inhibitor to allow time for the temperature-induced decay to be complete. cdc284, cdc28-13, and congenic wild-type cells were arrested in S phase by incubation with hydroxyurea at 230C. fter 4 hr, most of the mutant and wild-type cells had become synchronized, as determined by an increase in budded cells to almost 100% and by flow cytometric analysis and morphological observation (Fig. 1). t this time cells were shifted to 380C and incubated in the continued presence of hydroxyurea for 1 hr. Hydroxyurea was then removed by filtration with prewarmed medium and incubation was continued at 380C in the absence of hydroxurea. fter removal of hydroxyurea, wild-type cells recovered and resumed cycling as shown by the appearance of a steady-state fraction of unbudded cells and a dramatic increase in cell number (Fig. 1 and B). cdc28-13 mutant cells, after a lag, completed mitosis and became arrested in G1, as shown by an approximate doubling of cell number and the accumulation of unbudded cells with a in complement of nuclear DN (Fig. 1 -C). This result is consistent with previous observations as well as the proposal that the Cdc28 protein kinase does not have an essential mitotic function. cdc284 mutant cells, on the other hand, remained budded during the same time course, showed no cell number increase, and accumulated with a 2n complement of nuclear DN (Fig. 1 -C). Examination of the arrested cdc284 mutant cells stained with propidium iodide by fluorescence microscopy revealed that most have a nonextended nucleus in or near the bud aperture. This morphological phenotype is indicative of arrest in G2 (Fig. 1E). Viability of these arrested cells was confirmed by shifting the temperature to 230C at the end of the time course shown. Greater than 80% of the cells were then able to complete division (data not shown). Parallel observation of the arrested cdc28-13 mutant cells stained with propidium iodide revealed morphological and nuclear phenotypes associated with arrest in G1 (Fig. 1D). Specifically, the cells were unbudded and had asymmetric cytoplasmic projections, and the nucleus was centrally located. Thus of the two cdc28 mutations tested, only one revealed an essential function for the encoded protein kinase between the end of S phase and mitosis. The other, although conferring a lag in completion of the cell cycle relative to wild type, was not sufficiently defective in G2/M-phase functions to cause arrest in G2. These results indicate a G2 function for the Cdc28 protein kinase executed at some point during or after S phase of the cell cycle. Execution Point for the G2 Phase Cdc28 Function Is in G2. To localize the execution point of the Cdc28 protein kinase mitotic function, cdc28 mutant and wild-type cells were first arrested in G2 mutationally. Cells containing a deletion of the gene MIHJ, which encodes a mitotic inducer, and a copy of the Sc. pombe wee) gene, which encodes a mitotic inhibitor, expressed under control of the regulable GL) promoter, arrest in G2 when grown in galactose (40). When significant fractions of mutant and wild-type populations were arrested in G2 in this manner, as determined by morphological observation (data not shown), and then shifted to-restrictive temperature (380C) followed by release from the G2 block by filtration into glucose-containing medium, the wild-type cells were able to complete mitosis and initiated additional cell cycles (Fig. 2B). The cdc28-13 mutant, in this case, was unable to complete mitosis at 380C, as determined by cell

3 Genetics: Reed and Wittenberg I0t SC cdc28-1 3, 0 hr cdc28-4,,0 hr 80 I * WT * cdc cdc28 13 B,,--,-,cdc28-13, 5 hr cdc28-4, 11 E = 2 I' cci t t t HU 380 -HU Hours - l.-o 8 Icdc28-13, 8 hr Proc. Natl. cad. Sci. US 87 (1990) 5699, 5 hr Jcdc28-4, 8 hr FIG. 1. Cdc28 function relative to an S-phase block. Cells were arrested in S phase with hydroxyurea, shifted to 380C for 1 hr, and then released from the S-phase block. () Percentage of budded cells in the population as a function of time. (B) Cell number as a function of time. Points of addition of hydroxyurea (HU), shift to 380C, and removal of hydroxyurea (-HU) are indicated. *, Wild type (WT); *, cdc28-4;, cdc (C) Flow cytometric analysis of cdc28-13 (profiles on left) and cdc28-4 (profiles on right) mutant populations as a function of time. Times are 0 hr (top profiles); 5 hr (middle profiles), and 8 hr (bottom profiles). The ordinate for each plot represents frequency while the abcissa represents relative fluorescence. t the beginning time point (0 hr), the left peak represents cells with a in content of DN (G1 cells) and the right peak represents cells with a 2n content of DN (G2 + M phase cells). Note that there is a rightward drift (increased fluorescence) during each time course as a result of cell enlargement and cytoplasmic autofluorescence. (D) Fluorescent/phase-contrast micrograph of propidium iodide-stained cdc28-13 mutant at 8 hr. (E) Fluorescent/phase-contrast micrograph of propidium iodide-stained cdc28-4 mutant at 8 hr. (x350.) number, flow cytometry, and cellular morphology (Fig. 2 -C and E). These cells were viable, however, after release from the wee) IMIHI block at 380C, as demonstrated by their ability to complete mitosis and arrest in G, after a shift down to 360C (Fig. 2 -D). Presumably, the cdc28-13 mutation can establish a G2 block at 380C in this experiment because mutational perturbations in the wee)imihi regulatory network attenuate Cdc28 function in G2. The inability of these cdc28 mutant cells to recover at restrictive temperature from a previously established G2 block localizes the execution point for the Cdc28 mitotic function to G2. 80 t tt +gal 38% +glu Hours C 0 hr MV33.5 hr 7.5 hr 7.5 hr (36') H1 Kinase ctivity Through the Yeast Cell Cycle. In many cell types, the noncore histone H1 becomes phosphorylated during mitosis. This reversible phosphorylation has been associated with chromosome condensation during the early stages of mitosis. The protein kinase implicated in mitotic phosphorylation of histone H1 is the maturation-promoting factor complex containing cyclins and p34, the product of the CDC28/cdc2 homologs in these species (5-14). Similarly, the Cdc28 protein kinase in Sa. cerevisiae has been shown to be capable of efficiently phosphorylating histone H1 from animal sources (15), although the presence of a histone H1 FIG. 2. Cdc28 function relative to a G2 block. Cells were arrested mutationally in the G2 phase of the cell cycle. rrest was initiated at 0 hr by addition of galactose followed by elevation to 380C after 3.5 hr and filtration into preheated glucose-based medium at 4.0 hr. Subsequent incubation was at 380C or 360C. () Percent budded cells as a function of time. (B) Cell number as a function of time. Times of galactose addition (0 hr), shift to 380C (3.5 hr), and shift to glucose (4.0 hr) are indicated. *, Wild type (WT); m, cdc28-13 at 380C;, cdc28-13 at 360C. (C) Flow cytometric analysis of cdc28-13 mutant. Times are 0 hr (top profile), 3.5 hr (second profile), 7.5 hr with final incubation at 380C (third profile), and 7.5 hr with final incubation at 360C (bottom profile). Flow cytometric data are as in Fig. 1. (D) Fluorescent/phase-contrast micrograph of propidium iodide stained cdc28-13 mutant at 7.5 hr (360C). (E) Fluorescent/phase-contrast micrograph of propidium iodide-stained cdc28-13 mutant at 7.5 hr (38 C). (x480.)

4 5700 Genetics: Reed and Wittenberg homolog in yeast remains to be conclusively demonstrated. Nevertheless, we considered the ability to phosphorylate histone H1 a reasonable measure of G2/M activity of the Cdc28 protein kinase. Yeast cells of mating type a were presynchronized by incubation with the mating pheromone a factor. fter removal of mating pheromone by filtration, cells progress through at least one cell cycle with a high degree of synchrony. t regular intervals subsequent to release from the mating pheromone block, cells were harvested, lysates were prepared, and Cdc28 activity was assayed using bovine histone H1 as a substrate. s can be seen in Fig. 3, Cdc28-dependent histone H1 kinase activity was periodic and accumulated in the latter part of the cell cycle. Based on analysis of cell and nuclear morphologies (Fig. 3), the peak of protein kinase activity corresponded best to the fractions enriched for mitotic cells. Because the synchrony had decayed somewhat by the end of the first cell cycle, this experiment did not have sufficient resolution to determine precisely when histone H1 kinase activity was inactivated. To investigate the timing of histone H1 kinase activation, a population of growing cells was fractionated by centrifugal elutriation to obtain newly released G1 cells. Selection for small size gave a relatively pure population of daughter cells in G1. When assayed for histone H1 kinase activity, extracts from these cells showed relatively little activity (Fig. 3B). Thus by the time cells have completed mitosis and entered G1, the Cdc28 protein kinase has been inactivated. Extracts from elutriated fractions containing a heterogeneous population of cells that have progressed further through the cell cycle, as determined by size and degree of budding, showed high levels of histone kinase activity (Fig. 3B). U) =..-P Histone H1{ B Histone H1 Time (min) 12v 1 2 FIG. 3. Histone H1 kinase activity through the cell cycle. () Cells synchronized by mating pheromone block and release were assayed for Cdc28 protein kinase activity using histone H1 as a substrate as a function of progression through the cell cycle. The percentages of cells with small buds (e), mitotic nuclei (o), and no buds (n) are indicated as parameters of synchrony and cell cycle progression. Labeled histone H1 is indicated after SDS/PGE of reaction products and autoradiography. (B) Recently abcised daughter cells and a heterogeneous population of cells prepared by centrifugal elutriation were assayed for Cdc28 protein kinase activity, as described in. Lanes: 1, newly released daughter-enriched population; 2, mixed population. Proc. Natl. cad Sci. US 87 (1990) The substrate activity of an endogenous protein, p40, that coimmunoprecipitates with Cdc28 protein kinase complexes from asynchronous cellular lysates has been reported (25). By using synchronous cycling cultures, complexes capable of phosphorylating p40 in vitro were enriched in early G1 fractions. However, we now believe that result to be the consequence of in vitro mixing of free p40, which accumulates in G1, and mitotic Cdc28 complexes that we show here to be extremely active and that will inevitably contaminate G1 fractions because of the short interval between mitosis and G1. Thus, as synchrony improves, the window of p40 phosphorylation becomes increasingly restricted. The cell cycle profile of p40 phosphorylation based on a highly synchronous culture can be seen in Fig. 3, where p40 is the unmarked species of slightly lower mobility than that of histone H1. DISCUSSION Mitotic Role for Cdc28. It has long been perplexing that the Cdc28 protein kinase, clearly a structural homolog of the mitotic inducers for a broad spectrum of eukaryotic organisms, did not have a clearly demonstrable mitotic role (1, 2). It was furthermore shown that the CDC28 gene could rescue the G1 and mitotic defects conferred by a mutation in the homologous gene, cdc2', of Sc. pombe (34). This suggested that the Cdc28 protein kinase could assume a mitotic role when expressed in other organisms. One report (32) claimed that the Cdc28 protein kinase was required to proceed through mitosis in Sa. cerevisiae. This conclusion was based on an allele, cdc28-jn, which was reported to confer cell cycle arrest only in G2 and the demonstration that another allele, cdc28-6 (2), which normally confers arrest in G1, could arrest cells late in the cell cycle if they had been synchronized in G1. However, the execution point for this cell cycle block was determined to be in G1 (32). Our attempts to reproduce these experiments have been unsuccessful. We found that cdc28-jn mutant cells grow poorly at all temperatures but eventually arrest in G1 at the restrictive temperature, although not necessarily on the first cell cycle (M. D. Mendenhall and S.I.R., unpublished data). Our attempts to produce G2 arrest in other cdc28 mutants after release from a mating pheromone block were similarly unsuccessful. Flow cytometric analysis indicated that the arrest was heterogeneous relative to cell cycle position and transient in duration. Furthermore, a similar transient arrest was observed with wild-type cells, probably a result of heat shock during a sensitive window of the cell cycle (M. D. Mendenhall and S.I.R., unpublished data). The results reported here clearly establish a mitotic role for the Cdc28 protein kinase. cdc28 mutant cells, if prearrested in S or G2 phase, can be homogeneously synchronized in G2 by a shift to a high restrictive temperature (38 C). However, an asynchronous culture shifted to the same restrictive temperature arrests with >80% of the population in G1 (data not shown). There are a number of trivial explanations for this apparent paradox, most of which are probably related to properties intrinsic to temperature-sensitive mutant proteins. (i) Decay of the activity of a thermolabile protein most likely takes time. Hence we began the restrictive temperature incubation 1 hr prior to removal of the S-phase arresting agent. (ii) The G1 to S phase transition controlled by the Cdc28 protein kinase may require a higher level of activity than the G2 to M phase transition, explaining the bias toward arrest in G1 as the thermolabile function decays. We therefore used the maximum restrictive temperature tolerated in our genetic background without disturbance of growth and division, 38TC. It is noteworthy that most previous work with the alleles cdc284 and cdc28-13 has been at the minimum restrictive temperature, 34-36TC. t these temperatures, arrest in 02 is not observed for either allele using our protocol,

5 Genetics: Reed and Wittenberg suggesting that enough residual activity remains for the mitotic function even though the cells cannot proceed through G1 to S phase. This is undoubtedly the reason that the cdc28-13 allele arrested in G1 after completing mitosis even at 380C (Fig. 1). The completion of mitosis, however, did lag relative to the wild-type controls in that experiment (Fig. 1) and was blocked completely in the subsequent experiment (Fig. 2), indicating that reduced Cdc28 protein kinase activity conferred by the cdc28-13 allele has an impact on mitosis. In the fission yeast, Sc. pombe, the homologous gene cdc2' has execution points in G1 and G2 (3). Temperature-sensitive alleles do not appear to show an arrest-point bias at any restrictive temperature, although distinguishing G1 arrest from G2 arrest in fission yeast is technically more difficult than in budding yeast. It is likely that these differences reflect quantitative aspects of protein kinase function rather than qualitative differences. Thus, it appears that, for all their obvious dissimilarity, these highly diverged yeasts share a common central organization to their cell cycles based on a single protein kinase. The mihl/weel G2 arrest used in our experiments to localize the mitotic Cdc28 execution point is analogous to a similar mutational arrest in Sc. pombe mediated by the cdc25 and wee] mutations (40,42,43). In addition, tyrosine phosphorylation of the cdc2 protein kinase associated with this G2 arrest (22) also occurs under parallel circumstances with the Cdc28 protein kinase (unpublished observations). Our findings provide a landmark for the cell cycle of budding yeast. Prior investigations have proposed that mitosis in Sa. cerevisiae is controlled and begins prior to S phase (32). This view has been based on the previous difficulty in defining a G2 function for CDC28 in budding yeast and on cytological grounds-that a spindle-like structure is observed during most of the cell cycle (44). Our demonstration of a role for CDC28 after S phase indicates that mitosis in budding yeast is controlled late in the cell cycle, as is the case with most other eukaryotes. Given the particular structural features of the budding yeast mitotic apparatus, it is possible that mitosis is regulated at two points, with aspects related to spindle assembly controlled early and control of mitotic segregation of chromosomes occurring late. Timing of the Cdc28 Protein Kinase Function. Based on the experiments presented here, it can be said that CDC28 must function in G2 for cells to enter into mitosis. G1 function for the kinase has also been defined genetically (1, 2). When assayed in synchronous populations of cells using histone H1 as a substrate, Cdc28 protein kinase activity accumulates to high levels during the G2/M phase interval, corresponding to the genetically determined mitotic function. Only low levels of activity appear in G1. This is consistent with our observation that complexes containing the Cdc28 protein kinase and G1 cyclins, which presumably constitute the active G1 kinase, have low but detectable histone H1 kinase activity (45). Thus most of the histone kinase activity observed corresponds to the mitotic forms that presumably contain mitotic cyclins. The genes encoding such cyclins have recently been identified in Sa. cerevisiae (H. Richardson and S.I.R., unpublished data). Furthermore, the observation that histone H1 kinase activity is low in newly released postmitotic cells is consistent with the demonstration in other systems that maturation-promoting factor function is not required after metaphase (6, 8). In these respects the requirements for and observations of Cdc28 kinase activity parallel those defined for homologs in other organisms. We acknowledge Julian Thomas for expert technical assistance and Drs. Gary Cole, Helena Richardson, David Stone, Constance Stueland, and Katsunori Sugimoto, as well as Miguel de Barros Proc. Natl. cad. Sci. US 87 (1990) 5701 Lopes, for helpful discussions. This research was supported by U.S. Public Health Service Grant GM38328 to S.I.R. 1. Hartwell, L. H., Culotti, J., Pringle, J. & Reid, B. (1974) Science 183, Reed, S. I. (1980) Genetics 95, Nurse, P. & Bissett, Y. (1981) Nature (London) 292, Shuster, E. 0. & Byers, B. (1989) Genetics 123, Dunphy, W. G. & Newport, J. W. (1989) Cell 58, Murray,. W. & Kirschner, M. W. (1989) Nature (London) 339, Draetta, G. & Beach, D. (1988) Cell 54, Masui, Y. & Markert, C. L. (1971) J. Exp. Zool. 177, Gautier, J., Norbury, C., Lohka, M., Nurse, P. & Maller, J. (1988) Cell 54, Dunphy, W. G., Brizuela, L., Beach, D. & Newport, J. (1988) Cell 54, Labbe, J. C., Picard,., Peaucellier, G., Cavadore, J., Nurse, P. & Doree, M. (1989) Cell 57, rion, D., Meijer, L., Brizuela, L. & Beach, D. (1988) Cell 55, Draetta, G., Luca, F., Westerdorf, J., Brizuela, L., Ruderman, J. & Beach, D. (1989) Cell 56, Pines, J. & Hunter, T. (1989) Cell 58, Wittenberg, C. & Reed, S. I. (1988) Cell 54, Labbe, J.-C., Capony, J.-P., Caput, D., Cavadore, J.-C., Derancourt, J., Kaghad, M., Lelias, J.-M., Picard,. & Doree, M. (1989) EMBO J. 8, Booher, R. N., lfa, C. E., Hyams, J. S. & Beach, D. (1989) Cell 58, Moreno, S., Hayles, J. & Nurse, P. (1989) Cell 58, Gautier, J., Matsukawa, T., Nurse, P. & Maller, J. (1989) Nature (London) 339, Morla,. O., Draetta, G., Beach, D. & Wang, J. Y. J. (1989) Cell 58, Draetta, G., Piwnica-Worms, H., Morrison, D., Drucker, B., Roberts, T. & Beach, D. (1988) Nature (London) 336, Gould, K. L. & Nurse, P. (1989) Nature (London) 342, Murray,. W. & Kirschner, M. W. (1989) Nature (London) 339, Luca, F. C. & Ruderman, J. V. (1989) J. Cell Biol. 109, Mendenhall, M. D., Jones, C.. & Reed, S. I. (1987) Cell 50, Cross, F. (1988) Mol. Cell. Biol. 8, Nash, R., Tokiwa, G., nand, S., Erickson, K. & Futcher,. B. (1988) EMBO J. 7, Hadwiger, J.., Wittenberg, C., Richardson, H. E., de Barros Lopes, M. & Reed, S. I. (1989) Proc. Natl. cad. Sci. US 86, Richardson, H. E., Wittenberg, C., Cross, F. & Reed, S. I. (1989) Cell 59, Wittenberg, C. & Reed, S. I. (1989) Mol. Cell. Biol. 9, Lee, M. G. & Nurse, P. (1987) Nature (London) 327, Piggott, J. R., Rai, R. & Carter, B. L.. (1982) Nature (London) 298, Langan, T.., Gautier, J., Lohka, M., Hollingsworth, R., Moreno, S., Nurse, P., Maller, J. & Sclafani, R.. (1989) Mol. Cell. Biol. 9, Beach, D., Durkacz, B. & Nurse, P. (1982) Nature (London) 300, Reed, S. I., Hadwiger, J.. & Lorincz,. T. (1985) Proc. Natl. cad. Sci. US 82, Orr-Weaver, T. L., Szostak, J. W. & Rothstein, R. J. (1983) Methods Enzymol. 101, Rothstein, R. J. (1983) Methods Enzymol. 101, Sherman, F., Fink, G. & Hicks, J. B. (1982) Methods in Yeast Genetics (Cold Spring Harbor Lab., Cold Spring Harbor, NY). 39. Slater, M. L. (1973) J. Bacteriol. 113, Russell, P., Moreno, S. & Reed, S. I. (1989) Cell 57, Hutter, K.-J. & Eipel, H. E. (1979) J. Gen. Microbiol. 113, Russell, P. & Nurse, P. (1986) Cell 45, Russell, P. & Nurse, P. (1987) Cell 49, Byers, B. & Goetsch, L. (1975) J. Bacteriol. 124, Wittenberg, C., Sugimoto, K. & Reed, S. I. (1990) Cell 62, in press.